EP247 - Pretreatment ctDNA is Associated with Treatment Failure but not Clinical Response in Locally Advanced Rectal Cancer

Purpose/Background: The current standard of care for management of locally advanced rectal cancer is total neoadjuvant therapy (TNT) consisting of systemic chemotherapy and chemoradiotherapy. Patients determined to have a complete clinical response (cCR) or near complete clinical response (nCR) after TNT have the option of proceeding with nonoperative management, better known as "watch and wait." Patients with progressive disease (PD) or an incomplete response (IR) may require surgical resection. Stratification of patients at risk of early progression could facilitate identification of optimal treatment sequencing with selection of induction versus consolidation chemotherapy when initiating treatment or prevent potentially futile surgical intervention after TNT. Circulating tumor DNA (ctDNA) has been shown to correlate with disease recurrence in locally advanced rectal cancer. We aimed to evaluate the potential of pretreatment ctDNA value as an early predictor of treatment response or failure.

Methods/Interventions: We conducted a single-institution, retrospective analysis of adult patients with locally advanced rectal cancer who received TNT with ctDNA monitoring at a tertiary care center between June 2021 and September 2025. The primary objective was to evaluate if pretreatment ctDNA values could predict overall clinical response to TNT. The secondary objective was to evaluate the association between pretreatment ctDNA and treatment failure. Treatment failure was defined as PD during NAT or evidence of metastasis within six months of definitive treatment (TNT with watch and wait or TNT with surgery). Differences in ctDNA between CCR/nCR and IR/PD were assessed with nonparametric testing, and discriminatory performance was evaluated with ROC analysis. The association between pretreatment ctDNA and time to treatment failure was examined with a univariate Cox proportional hazards model.

Results/Outcomes: Thirty-one patients were included in our analysis, and six patients (19.4%) met criteria for a treatment failure. There was no significant difference in clinical response across the distribution of pretreatment ctDNA values (Wilcoxon p = 0.44; Figure 1), and ROC analysis showed limited discrimination between treatment response groups (AUC = 0.58; Figure 2). However, higher pretreatment ctDNA values were significantly associated with treatment failure. Each 1-unit increase in ctDNA corresponded to a 9% higher risk of treatment failure (HR = 1.09, 95% Wald CI 0.99-1.20; likelihood ratio test p = 0.002).

Conclusion/Discussion: In patients with locally advanced rectal cancer undergoing TNT, higher pretreatment ctDNA levels were associated with an increased risk of treatment failure; however, pretreatment ctDNA values were not predictive of overall clinical response. These data suggest a potential role for pretreatment ctDNA in identifying patients at risk of disease progression with possible implications for treatment optimization in rectal cancer. Future work should include larger prospective trials to further evaluate the predictive role of ctDNA and its clinical implications.

Figure 1. Distribution of pretreatment ctDNA by NAT response status
Violin plot showing the distribution of pretreatment ctDNA levels across clinical response groups following total neoadjuvant therapy. No significant difference was observed between groups (Wilcoxon p = 0.44), indicating limited discriminatory value of ctDNA for predicting treatment response.
Figure 2. ROC curve for pretreatment ctDNA and NAT response
Receiver operating characteristic (ROC) curve evaluating the ability of pretreatment ctDNA values to discriminate between treatment response (CCR/nCR and IR/PD) following total neoadjuvant therapy. The area under the curve (AUC) was 0.58, indicating limited discriminatory performance.